Using a spoken corpus compiled from subtitles to study apologies

Apologies, like other speech acts, are difficult to study, and they are studied in a variety of ways, including Discourse Completion Tests, role plays, gathering of natural language, and using spoken corpora. Each method has advantages and disadvantages. One possibility for developing a spoken corpus is using subtitles downloaded from the DVDs of movies or television programs. In this study, I have used the first two seasons of a US television comedy called Modern Family to develop a corpus of spoken English. Using five keywords ― sorry, excuse, apologize, forgive, and pardon ― I identified the apologies found in the corpus and analyzed them, with particular attention to the types of apologies that would be difficult or impossible to elicit using a Discourse Completion Test.